INDEX

Explanations

names in a non-English language

oai_token-act-pair · gpt-3.5-turbo

punctuation, particularly commas, in the text

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

 footing

-0.64

FIX

-0.60

 reasons

-0.60

ird

-0.58

enhagen

-0.57

pport

-0.56

 discretionary

-0.56

 exha

-0.55

 investment

-0.55

 decisions

-0.55

POSITIVE LOGITS

aka

1.00

huh

0.91

etc

0.86

 respectively

0.81

 meanwhile

0.77

 which

0.76

 wherein

0.76

 although

0.75

Jr

0.74

Downloadha

0.74

Activations Density 1.018%

names in a non-English language

punctuation, particularly commas, in the text

No Comments

No Known Activations