INDEX

Explanations

This neuron is primarily looking for names of people and places in various contexts

oai_token-act-pair · gpt-3.5-turbo

phrases indicating significant events or notable actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

itably

-0.79

ancial

-0.75

escription

-0.74

;;;;;;;;;;;;

-0.71

irlf

-0.71

uilt

-0.70

ividual

-0.69

ividually

-0.66

retty

-0.66

lished

-0.66

POSITIVE LOGITS

CLASSIFIED

1.09

NetMessage

0.95

WAYS

0.89

 behavi

0.87

natureconservancy

0.81

GoldMagikarp

0.81

 largeDownload

0.81

 welf

0.76

 toile

0.76

Interstitial

0.76

Activations Density 6.087%

This neuron is primarily looking for names of people and places in various contexts

phrases indicating significant events or notable actions

No Comments

No Known Activations