INDEX

Explanations

mentions of the city "Tucson"

oai_token-act-pair · gpt-3.5-turbo

references to specific elements related to health conditions and locations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

train

-0.83

phal

-0.75

Cand

-0.75

å£«

-0.75

phe

-0.74

Versions

-0.71

vous

-0.71

heads

-0.70

fol

-0.67

handler

-0.67

POSITIVE LOGITS

olkien

1.22

ribute

1.02

olerance

1.01

ango

0.97

oday

0.95

akedown

0.95

urtle

0.95

uple

0.93

oxin

0.93

ractor

0.92

Activations Density 0.037%

mentions of the city "Tucson"

references to specific elements related to health conditions and locations

No Comments

No Known Activations