INDEX

Explanations

instances of lying or deception

oai_token-act-pair · gpt-3.5-turbo

instances of deception or dishonesty, particularly involving lying to authorities or the public

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

eric

-0.72

winner

-0.67

pour

-0.67

 specialization

-0.63

 guiActiveUnfocused

-0.62

Interstitial

-0.62

pora

-0.62

 Radius

-0.62

 procession

-0.62

ateur

-0.60

POSITIVE LOGITS

 omission

1.03

 deceive

0.81

 falsely

0.79

uth

0.79

 perjury

0.78

 testifying

0.78

 dece

0.77

 incrim

0.77

truth

0.73

Activations Density 0.287%

instances of lying or deception

instances of deception or dishonesty, particularly involving lying to authorities or the public

No Comments

No Known Activations