INDEX

Explanations

mentions of the word 'pen'

oai_token-act-pair · gpt-3.5-turbo

mentions of the word "pen."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ndum

-0.82

 Lumpur

-0.75

ONEY

-0.73

ONSORED

-0.73

IELD

-0.69

OHN

-0.66

IAS

-0.64

DEF

-0.64

UTH

-0.64

 Influence

-0.63

POSITIVE LOGITS

ultimate

1.43

itent

1.40

manship

1.09

pen

1.04

ning

1.02

insula

1.01

cil

0.99

cill

0.98

elope

0.97

nington

0.89

Activations Density 0.007%

mentions of the word 'pen'

mentions of the word "pen."

No Comments

No Known Activations