INDEX

Explanations

medical terms or symptoms

oai_token-act-pair · gpt-3.5-turbo

terms indicating symptoms of issues or conditions

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rences

-0.73

 upstream

-0.69

inals

-0.67

 proc

-0.66

 glam

-0.64

ortal

-0.64

 conclusive

-0.64

cov

-0.62

 inval

-0.62

pil

-0.60

POSITIVE LOGITS

Wallet

0.77

Marco

0.77

DEM

0.77

ãĥīãĥ©

0.76

urai

0.75

Args

0.73

Ari

0.71

rek

0.71

axis

0.70

Fine

0.69

Activations Density 0.000%

medical terms or symptoms

terms indicating symptoms of issues or conditions

No Comments

No Known Activations