INDEX

Explanations

periods

oai_token-act-pair · gpt-3.5-turbo

periods or full stops in the text

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

alyses

-0.76

eworld

-0.73

zsche

-0.69

opter

-0.69

onyms

-0.69

oven

-0.66

 millennia

-0.66

ankind

-0.66

sworth

-0.64

 spac

-0.64

POSITIVE LOGITS

icio

0.88

Malley

0.87

 Bernard

0.83

 Bruce

0.82

Cu

0.81

 Andrew

0.81

 Gavin

0.80

 Gerald

0.78

 Scott

0.78

 Maurice

0.78

Activations Density 0.026%

periods

periods or full stops in the text

No Comments

No Known Activations