INDEX

Explanations

phrases related to intervention or action

oai_token-act-pair · gpt-3.5-turbo

phrases indicating involvement or participation in an action or event

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 Ming

-0.62

haus

-0.62

ãĥ´

-0.62

achus

-0.61

 headache

-0.60

omnia

-0.59

 occurrence

-0.58

eming

-0.58

 Corpus

-0.58

worth

-0.57

POSITIVE LOGITS

 stride

0.88

 bounds

0.85

circle

0.83

 sidx

0.75

ETHOD

0.75

agall

0.74

 srfAttach

0.73

steps

0.69

 sideways

0.69

 paddle

0.69

Activations Density 0.058%

phrases related to intervention or action

phrases indicating involvement or participation in an action or event

No Comments

No Known Activations