INDEX

Explanations

phrases related to past events or actions

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

 favour

-0.76

adr

-0.74

rouse

-0.73

venge

-0.66

 favor

-0.65

icket

-0.63

ocracy

-0.63

 whichever

-0.62

iger

-0.61

abiding

-0.60

POSITIVE LOGITS

noon

1.02

etheless

0.90

 editions

0.88

 versions

0.86

 Thoughts

0.86

Ago

0.85

 Days

0.78

 Releases

0.76

 Accounts

0.76

 Seasons

0.75

Activations Density 0.058%

phrases related to past events or actions

No Comments

No Known Activations

phrases related to past events or actions

No Comments

No Known Activations