INDEX

Explanations

references to specific individual entities

oai_token-act-pair · gpt-3.5-turbo

references to a central or significant individual in various contexts

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

inders

-0.84

ategory

-0.80

iminary

-0.73

pace

-0.69

actionDate

-0.68

rang

-0.65

avascript

-0.65

ICO

-0.64

ourn

-0.63

helm

-0.63

POSITIVE LOGITS

 Hundred

0.90

 hundred

0.81

 Thousand

0.76

liest

0.70

 eyed

0.70

 wonders

0.70

 sided

0.69

 thing

0.69

 thousand

0.66

 person

0.65

Activations Density 0.032%

references to specific individual entities

references to a central or significant individual in various contexts

No Comments

No Known Activations