INDEX

Explanations

phrases related to information provided by anonymous sources

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

neys

-0.82

union

-0.78

nant

-0.74

tons

-0.74

ney

-0.73

Done

-0.72

NEY

-0.71

rex

-0.69

yg

-0.68

abase

-0.66

POSITIVE LOGITS

 anonymity

1.11

ously

1.07

 anonym

1.01

onym

0.89

 anonymously

0.82

onyms

0.80

 pseudonym

0.74

 anonymous

0.74

arial

0.73

 informant

0.72

Activations Density 0.032%

phrases related to information provided by anonymous sources

No Comments

No Known Activations