INDEX

Explanations

comparisons of rankings or positions in a list

oai_token-act-pair · gpt-3.5-turbo

phrases indicating rankings or positions in comparison to others

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

beh

-0.72

Limited

-0.67

Register

-0.66

oresc

-0.65

leeve

-0.64

vari

-0.63

arant

-0.63

trial

-0.62

translation

-0.62

anu

-0.62

POSITIVE LOGITS

ansas

0.74

 rivals

0.72

 fellow

0.72

 Laksh

0.71

 Norway

0.68

 rival

0.68

 Avatar

0.67

 Coca

0.65

 Saud

0.65

 neighboring

0.65

Activations Density 0.168%

comparisons of rankings or positions in a list

phrases indicating rankings or positions in comparison to others

No Comments

No Known Activations