INDEX

Explanations

actions involving physical aggression or conflict

oai_token-act-pair · gpt-3.5-turbo

instances of the word "and."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

onymous

-0.82

éŃĶ

-0.78

:(

-0.76

:[

-0.74

Ĥİ

-0.73

ESE

-0.72

uthor

-0.70

Student

-0.70

BIL

-0.70

Commission

-0.69

POSITIVE LOGITS

 thus

0.92

 consequently

0.92

 therefore

0.88

 assorted

0.86

 hence

0.85

 then

0.83

 possibly

0.81

 flats

0.79

 vice

0.79

rogens

0.78

Activations Density 0.884%

actions involving physical aggression or conflict

instances of the word "and."

No Comments

No Known Activations