INDEX

Explanations

expert-related terms or phrases in the text

oai_token-act-pair · gpt-3.5-turbo

references to expert testimony and expertise

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

apo

-0.77

strip

-0.72

[|

-0.67

 Yamato

-0.65

oÄŁ

-0.63

FACE

-0.63

Bom

-0.60

 Mellon

-0.60

 Marse

-0.59

 externalToEVAOnly

-0.59

POSITIVE LOGITS

 consulted

1.07

 opinion

0.94

 advise

0.89

ise

0.88

udic

0.87

 witness

0.85

 advice

0.85

 reviewer

0.83

 specializing

0.83

sonian

0.82

Activations Density 0.057%

expert-related terms or phrases in the text

references to expert testimony and expertise

No Comments

No Known Activations