INDEX

Explanations

proper nouns or names of entities

oai_token-act-pair · gpt-3.5-turbo

phrases indicating a group or a collective entity

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 anytime

-0.69

ario

-0.65

rification

-0.61

ober

-0.60

 translates

-0.60

ysc

-0.59

culosis

-0.58

olation

-0.58

 introduction

-0.58

 commute

-0.57

POSITIVE LOGITS

 those

0.81

ĪĴ

0.73

Ĳ

0.72

 casualties

0.71

those

0.71

 whom

0.66

Sources

0.66

addons

0.66

 innumerable

0.66

stad

0.64

Activations Density 0.028%

proper nouns or names of entities

phrases indicating a group or a collective entity

No Comments

No Known Activations