INDEX

Explanations

mentions of corruption and related concepts

oai_token-act-pair · gpt-3.5-turbo

references to corruption 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 @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Able

-0.84

 Atkins

-0.70

 Tanner

-0.69

kees

-0.68

Wave

-0.66

 Instruct

-0.66

 Advent

-0.65

 spaced

-0.65

dule

-0.65

irds

-0.64

POSITIVE LOGITS

 corruption

3.84

 Corruption

2.94

corruption

2.74

 corrupt

2.37

 bribery

2.09

 corrupted

2.03

 graft

1.75

 bribes

1.59

 bribe

1.44

 criminality

1.36

Activations Density 0.036%

mentions of corruption and related concepts

references to corruption in various contexts

No Comments

No Known Activations