INDEX

Explanations

punctuation marks

oai_token-act-pair · gpt-3.5-turbo

combinations or sequences of commas and related punctuation

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

terday

-0.74

ledged

-0.71

cedented

-0.62

bidden

-0.59

 Yesterday

-0.59

ŃĶ

-0.58

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.58

Alias

-0.58

ãĤ¨ãĥ«

-0.58

ophon

-0.56

POSITIVE LOGITS

 preferably

1.29

 then

1.05

 THEN

0.99

 rinse

0.94

 sprinkle

0.92

 lest

0.90

then

0.88

 arrange

0.86

 grab

0.86

 toss

0.85

Activations Density 0.229%

punctuation marks

combinations or sequences of commas and related punctuation

No Comments

No Known Activations