INDEX

Explanations

names of individuals, possibly with particular initials (such as W.F.) and surnames

oai_token-act-pair · gpt-3.5-turbo

references to notable individuals or historical figures

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

culosis

-0.69

vous

-0.66

ufact

-0.66

TVs

-0.65

alore

-0.64

basketball

-0.62

cffff

-0.62

incial

-0.61

Pg

-0.61

duino

-0.61

POSITIVE LOGITS

stad

0.85

aley

0.83

 Roberts

0.76

burn

0.73

gren

0.73

mann

0.70

hyde

0.70

enstein

0.69

III

0.67

agos

0.67

Activations Density 0.200%

names of individuals, possibly with particular initials (such as W.F.) and surnames

references to notable individuals or historical figures

No Comments

No Known Activations