INDEX

Explanations

phrases related to fire or combustion

oai_token-act-pair · gpt-3.5-turbo

references to fire or burning

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

odore

-0.74

ournal

-0.73

onsense

-0.72

xus

-0.71

 reluct

-0.71

alian

-0.70

awaru

-0.69

udeau

-0.68

ordan

-0.68

ocre

-0.67

POSITIVE LOGITS

 burning

1.05

ished

1.04

ishing

1.02

 burn

1.02

burning

0.93

 burns

0.93

ishes

0.92

 hotter

0.91

 burned

0.85

houses

0.85

Activations Density 0.025%

phrases related to fire or combustion

references to fire or burning

No Comments

No Known Activations