INDEX

Explanations

mentions of unethical behavior or wrongdoings

oai_token-act-pair · gpt-3.5-turbo

references to misconduct 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 @ 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

izen

-0.76

stanbul

-0.76

iku

-0.68

nee

-0.67

 Zuckerberg

-0.67

etically

-0.66

kos

-0.66

frey

-0.64

pop

-0.64

imb

-0.64

POSITIVE LOGITS

 misconduct

1.16

onduct

1.01

 allegations

0.92

 unfocusedRange

0.82

 malf

0.80

 disclosures

0.77

 abuses

0.75

 scandals

0.75

HAEL

0.74

acies

0.74

Activations Density 0.007%

mentions of unethical behavior or wrongdoings

references to misconduct in various contexts

No Comments

No Known Activations