INDEX

Explanations

proper nouns, specifically names of places or people

oai_token-act-pair · gpt-3.5-turbo

references to specific individuals or notable figures

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

aciously

-0.86

bsp

-0.76

 Gutenberg

-0.75

uous

-0.71

EMP

-0.69

information

-0.68

 debtor

-0.68

 doct

-0.67

unction

-0.66

ifier

-0.65

POSITIVE LOGITS

 Carroll

1.09

sburg

0.90

mont

0.88

ton

0.86

 Gardens

0.84

lean

0.79

 Shelby

0.79

oqu

0.78

dale

0.75

anut

0.75

Activations Density 0.010%

proper nouns, specifically names of places or people

references to specific individuals or notable figures

No Known Activations