INDEX

Explanations

phrases or names containing "le" followed by a number

oai_token-act-pair · gpt-3.5-turbo

instances of the word "le" in different contexts

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

rador

-0.99

 Seym

-0.99

raints

-0.97

yrinth

-0.91

iaries

-0.89

ĸļ

-0.88

ividual

-0.86

 carbohyd

-0.85

raint

-0.85

inelli

-0.83

POSITIVE LOGITS

ttes

1.20

lla

1.17

agues

1.16

phant

0.96

isure

0.96

utenant

0.89

asure

0.87

lean

0.86

eping

0.85

phrine

0.85

Activations Density 0.026%

phrases or names containing "le" followed by a number

instances of the word "le" in different contexts

No Comments

No Known Activations