INDEX

Explanations

mentions of medical personnel and procedures

oai_token-act-pair · gpt-3.5-turbo

terms related to medical concepts or healthcare situations

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

DOM

-0.81

ACK

-0.76

 indign

-0.71

TRUMP

-0.70

hyde

-0.66

WT

-0.66

 Argon

-0.64

 Klan

-0.63

 Spencer

-0.63

âĻ¦

-0.63

POSITIVE LOGITS

med

1.09

allion

0.99

ications

0.96

iterranean

0.94

itations

0.92

icum

0.89

undai

0.89

ulla

0.88

ics

0.87

itative

0.86

Activations Density 0.003%

mentions of medical personnel and procedures

terms related to medical concepts or healthcare situations

No Comments

No Known Activations