INDEX

Explanations

decisions or judgments made by an authority figure

oai_token-act-pair · gpt-3.5-turbo

instances of the word "ruled."

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

issance

-0.78

akra

-0.76

ription

-0.75

rity

-0.71

ionage

-0.69

ickr

-0.67

ipal

-0.66

Newsletter

-0.64

ãĥ¤

-0.64

 Pilgrim

-0.64

POSITIVE LOGITS

phas

0.83

maker

0.76

inker

0.75

 unfit

0.70

tto

0.68

rooms

0.67

ULE

0.66

book

0.66

witz

0.65

sets

0.63

Activations Density 0.023%

decisions or judgments made by an authority figure

instances of the word "ruled."

No Comments

No Known Activations