INDEX

Explanations

medical-related terms and abbreviations

oai_token-act-pair · gpt-3.5-turbo

mentions of the acronym "DR"

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 nest

-0.71

 tight

-0.68

 Kaufman

-0.68

 Pluto

-0.67

 midway

-0.63

 closer

-0.63

 cottage

-0.63

Fut

-0.63

 wide

-0.62

 comet

-0.61

POSITIVE LOGITS

DR

4.32

DR

1.91

dr

1.72

DA

1.50

DP

1.44

DW

1.40

DE

1.37

DS

1.37

DL

1.34

DF

1.34

Activations Density 0.008%

medical-related terms and abbreviations

mentions of the acronym "DR"

No Comments

No Known Activations