INDEX

Explanations

phrases related to actions or events (e.g., failing to mention, making assertions)

oai_token-act-pair · gpt-3.5-turbo

phrases indicating omissions or statements of fact in reports

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 Celt

-0.69

 Syndrome

-0.68

 Transformation

-0.65

 Forsaken

-0.64

 dying

-0.64

 Ascension

-0.64

 contracted

-0.63

Xan

-0.63

 Acquisition

-0.63

wear

-0.62

POSITIVE LOGITS

atform

0.93

undown

0.82

arnaev

0.81

legraph

0.80

 outlines

0.80

 outlining

0.78

 outline

0.78

uggest

0.75

 summar

0.75

 evid

0.74

Activations Density 0.405%

phrases related to actions or events (e.g., failing to mention, making assertions)

phrases indicating omissions or statements of fact in reports

No Comments

No Known Activations