INDEX

Explanations

mentions or instances of honesty

oai_token-act-pair · gpt-3.5-turbo

instances of the word "honest" and variations of it in various contexts

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

oval

-0.68

Ĥİ

-0.67

 curfew

-0.67

 Warrant

-0.67

İĭ

-0.65

lav

-0.64

rior

-0.64

joined

-0.63

 Supplement

-0.62

 Krish

-0.62

POSITIVE LOGITS

 honest

0.93

 appraisal

0.86

Honest

0.81

 truthful

0.79

 honesty

0.77

spection

0.75

 broker

0.74

 mistakes

0.73

 candid

0.73

 Honest

0.73

Activations Density 0.054%

mentions or instances of honesty

instances of the word "honest" and variations of it in various contexts

No Comments

No Known Activations