INDEX

Explanations

phrases related to defending actions or individuals

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

hook

-0.68

hall

-0.68

FORE

-0.65

 explode

-0.65

 bleed

-0.63

ombies

-0.62

way

-0.62

çīĪ

-0.62

psy

-0.60

ptic

-0.59

POSITIVE LOGITS

 against

1.17

 Against

1.05

against

0.98

atively

0.87

 vigorously

0.84

iveness

0.81

ively

0.77

ously

0.75

antly

0.75

ably

0.75

Activations Density 0.037%

phrases related to defending actions or individuals

No Comments

No Known Activations