INDEX

Explanations

references to political figures or events, particularly related to security and controversies

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

 ende

-0.72

 incorpor

-0.71

 ponder

-0.71

 mushroom

-0.69

 imagination

-0.68

 controvers

-0.65

 retrieval

-0.65

 rundown

-0.64

 electorate

-0.64

 unwanted

-0.64

POSITIVE LOGITS

ï¸ı

1.23

ÃĽ

0.96

ï¸

0.93

Â¯

0.92

STEM

0.88

cue

0.86

Â°

0.84

said

0.83

âĢł

0.80

#$

0.80

Activations Density 0.795%

references to political figures or events, particularly related to security and controversies

No Comments

No Known Activations