INDEX

Explanations

phrases that indicate a lack of experience or having never done something before

oai_token-act-pair · gpt-3.5-turbo

statements expressing a lack of prior experience or knowledge about something

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

States

-0.79

hol

-0.65

Rap

-0.63

heimer

-0.62

MG

-0.61

 Pascal

-0.61

states

-0.60

mart

-0.60

 Heads

-0.59

Pe

-0.59

POSITIVE LOGITS

been

1.07

theless

0.94

 been

0.91

 undergone

0.83

ĸļ

0.79

 Been

0.78

 EVER

0.77

 tasted

0.77

izens

0.75

 seen

0.74

Activations Density 0.051%

phrases that indicate a lack of experience or having never done something before

statements expressing a lack of prior experience or knowledge about something

No Comments

No Known Activations