INDEX

Explanations

phrases relating to a particular aspect or concept, but the examples provided do not reveal a common theme

oai_token-act-pair · gpt-3.5-turbo

the term "thing," often referring to various subjects or concepts in discussion

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

inav

-0.90

incinn

-0.75

rylic

-0.72

ardi

-0.71

ervation

-0.71

osponsors

-0.69

oufl

-0.69

irl

-0.68

cling

-0.67

ctic

-0.67

POSITIVE LOGITS

Else

0.91

 thing

0.88

 happ

0.86

iverse

0.85

 Valiant

0.83

 happening

0.82

 happened

0.79

REDACTED

0.78

 Thing

0.77

worm

0.75

Activations Density 0.028%

phrases relating to a particular aspect or concept, but the examples provided do not reveal a common theme

the term "thing," often referring to various subjects or concepts in discussion

No Comments

No Known Activations