INDEX

Explanations

verbs that describe observing or noticing, as well as words indicating application or persistence

oai_token-act-pair · gpt-3.5-turbo

phrases indicating conditions or observations that are clear and applicable

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 wishes

-0.64

 foregoing

-0.63

inventoryQuantity

-0.61

 fame

-0.58

dule

-0.57

 preparation

-0.56

 scrap

-0.55

claimer

-0.55

wer

-0.55

 clocks

-0.54

POSITIVE LOGITS

 everywhere

1.05

 throughout

0.90

 elsewhere

0.90

 Nationwide

0.85

 across

0.84

ï¸ı

0.82

today

0.74

 irrespective

0.74

igmatic

0.73

 Everywhere

0.71

Activations Density 0.288%

verbs that describe observing or noticing, as well as words indicating application or persistence

phrases indicating conditions or observations that are clear and applicable

No Comments

No Known Activations