INDEX

Explanations

words related to honesty and dishonesty

oai_token-act-pair · gpt-3.5-turbo

terms related to honesty and dishonesty

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

LAN

-0.83

fleet

-0.82

Fan

-0.79

 Kard

-0.70

avia

-0.70

LOCK

-0.67

ergy

-0.66

ENTS

-0.66

 Mutant

-0.66

 Archdemon

-0.66

POSITIVE LOGITS

 honesty

0.94

cipled

0.94

esty

0.90

acies

0.85

 broker

0.83

 cheat

0.80

 cand

0.78

 dishonest

0.77

 cheated

0.74

uth

0.73

Activations Density 0.015%

words related to honesty and dishonesty

terms related to honesty and dishonesty

No Comments

No Known Activations