INDEX

Explanations

words related to mysterious or secret organizations or operations

oai_token-act-pair · gpt-3.5-turbo

references to specific locations or entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 âĢº

-0.69

 ..........

-0.66

ãĥĥ

-0.65

terday

-0.65

Ú

-0.62

ctica

-0.61

å¸

-0.59

Mb

-0.58

 âĵĺ

-0.57

 seism

-0.56

POSITIVE LOGITS

ultimate

0.83

hest

0.75

iest

0.71

onian

0.69

enium

0.68

ucer

0.68

pee

0.67

yssey

0.67

cients

0.67

verse

0.66

Activations Density 0.383%

words related to mysterious or secret organizations or operations

references to specific locations or entities

No Comments

No Known Activations