INDEX

Explanations

phrases related to evaluations or judgments

oai_token-act-pair · gpt-3.5-turbo

terms related to justification and validity in contexts involving decisions or statements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 domest

-0.74

 kindred

-0.68

 indo

-0.67

 oppressed

-0.64

etus

-0.64

 resil

-0.61

 decad

-0.60

ut

-0.60

 plent

-0.60

 everyday

-0.59

POSITIVE LOGITS

due

1.04

 owing

0.98

 because

0.93

due

0.93

because

0.91

 citing

0.82

 considering

0.81

 despite

0.79

 pending

0.78

 barring

0.77

Activations Density 0.416%

phrases related to evaluations or judgments

terms related to justification and validity in contexts involving decisions or statements

No Comments

No Known Activations