INDEX

Explanations

dates mentioned in the text

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

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

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rehensive

-0.43

rouse

-0.43

 cumbers

-0.42

hetically

-0.42

asive

-0.42

 withd

-0.41

 entertained

-0.41

 encount

-0.41

 enriched

-0.40

 volunte

-0.39

POSITIVE LOGITS

0.95

0.94

 October

0.90

 September

0.90

0.89

 July

0.89

 November

0.89

Activations Density 0.415%

dates mentioned in the text

No Comments

No Known Activations