INDEX

Explanations

the term "well-known."

oai_token-act-pair · gpt-3.5-turbo

the phrase "well known."

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

hyde

-0.87

hip

-0.81

atto

-0.71

illary

-0.70

rush

-0.69

adena

-0.69

amera

-0.66

ategory

-0.65

ongyang

-0.65

iferation

-0.64

POSITIVE LOGITS

enough

1.05

 enough

0.94

 suited

0.90

spring

0.82

wired

0.75

 Enough

0.75

 behaved

0.73

Known

0.73

baum

0.71

esley

0.68

Activations Density 0.042%

the term "well-known."

the phrase "well known."

No Comments

No Known Activations