INDEX

Explanations

bits of text or metadata related to titles and identifiers

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

names and other proper nouns

oai_token-act-pair · gpt-3.5-turbo

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

ãĥ¼ãĥĨ

-0.71

 behavi

-0.68

 thous

-0.64

toe

-0.64

 compar

-0.63

 bona

-0.61

 comparisons

-0.60

 competitor

-0.59

 competition

-0.59

 criteria

-0.59

POSITIVE LOGITS

arah

0.87

anta

0.86

hz

0.85

opa

0.85

ESSION

0.83

lt

0.82

oshi

0.81

df

0.81

bps

0.80

ayers

0.80

Activations Density 0.036%

bits of text or metadata related to titles and identifiers

names and other proper nouns

No Comments

No Known Activations