INDEX

Explanations

places or events with names that end in 'atra'

oai_token-act-pair · gpt-3.5-turbo

references to specific locations or cultural contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

cre

-0.68

 prob

-0.66

 pool

-0.65

 shades

-0.63

WR

-0.62

 confirmation

-0.62

 stud

-0.61

Inf

-0.61

Vol

-0.60

 Cong

-0.58

POSITIVE LOGITS

atra

4.96

atri

1.31

arta

1.18

ahar

1.05

andra

1.04

atism

1.02

asca

0.98

agra

0.97

peria

0.95

atana

0.95

Activations Density 0.004%

places or events with names that end in 'atra'

references to specific locations or cultural contexts

No Comments

No Known Activations