INDEX

Explanations

strong language related to actions or situations.

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

words related to emotional responses and reactions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

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

AMY

-0.66

umerable

-0.60

ationally

-0.57

ational

-0.53

Sph

-0.53

bourg

-0.52

bern

-0.52

 scope

-0.52

 Express

-0.52

umbers

-0.51

POSITIVE LOGITS

off

2.65

off

2.23

Off

2.03

OFF

1.90

Off

1.89

offs

1.86

OFF

1.78

 away

1.03

 offline

1.02

awa

0.91

Activations Density 0.736%

**strong language related to actions or situations.**

words related to emotional responses and reactions

No Comments

No Known Activations

strong language related to actions or situations.