INDEX

Explanations

words related to actions or allegations of wrongdoing or misconduct

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

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

etically

-0.73

etic

-0.70

cube

-0.69

NetMessage

-0.69

pop

-0.66

alg

-0.66

UNCH

-0.65

answer

-0.65

iku

-0.64

ERC

-0.64

POSITIVE LOGITS

 perpetrated

1.29

 wrongdoing

1.06

 misconduct

1.04

 abuses

1.04

 committed

1.03

 allegations

1.00

 violations

0.96

 crimes

0.92

 atrocities

0.91

 inflicted

0.91

Activations Density 0.113%

words related to actions or allegations of wrongdoing or misconduct

No Comments

No Known Activations