INDEX

Explanations

names of specific individuals

oai_token-act-pair · gpt-3.5-turbo

proper names, particularly those of individuals and related entities

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Marginal

-0.80

kt

-0.79

NetMessage

-0.76

EMENT

-0.74

istors

-0.69

RANT

-0.68

ansas

-0.65

Leader

-0.64

reens

-0.64

 Bundy

-0.63

POSITIVE LOGITS

 Chop

0.86

aum

0.82

ley

0.74

ħ

0.72

Dai

0.71

idge

0.71

 Calories

0.70

lesi

0.69

lett

0.68

les

0.67

Activations Density 0.016%

names of specific individuals

proper names, particularly those of individuals and related entities

No Comments

No Known Activations