INDEX

Explanations

names of individuals

oai_token-act-pair · gpt-3.5-turbo

characters or symbols that may appear in a written context, specifically focusing on certain letters or symbols

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 toget

-0.67

 succeeding

-0.65

vou

-0.65

 packed

-0.64

 bearing

-0.62

ãģ®éŃĶ

-0.61

 Authority

-0.60

QC

-0.60

 Shine

-0.60

 thirds

-0.59

POSITIVE LOGITS

ulhu

1.24

odan

1.05

ark

1.01

irt

0.98

irk

0.97

arn

0.97

arks

0.96

ork

0.96

audi

0.96

arna

0.95

Activations Density 0.067%

names of individuals

characters or symbols that may appear in a written context, specifically focusing on certain letters or symbols

No Comments

No Known Activations