INDEX

Explanations

mentions of dignity

oai_token-act-pair · gpt-3.5-turbo

references to the concept of dignity

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

WER

-0.75

Ack

-0.72

enegger

-0.67

 fielded

-0.65

eding

-0.65

VEN

-0.64

eder

-0.62

suggest

-0.62

addon

-0.62

vik

-0.61

POSITIVE LOGITS

 dignity

0.95

ility

0.94

ously

0.84

ilege

0.82

":""},{"

0.80

 Haram

0.77

acies

0.76

 bestowed

0.76

ileged

0.75

ity

0.75

Activations Density 0.028%

mentions of dignity

references to the concept of dignity

No Comments

No Known Activations