INDEX

Explanations

instances where a significant action or decision is being reported on or discussed

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

 sqor

-0.77

omial

-0.74

vae

-0.72

Condition

-0.66

 attm

-0.65

 Koran

-0.64

errors

-0.64

ordon

-0.62

aez

-0.62

inges

-0.62

POSITIVE LOGITS

 toward

0.97

backs

0.97

able

0.95

 towards

0.95

ments

0.90

over

0.81

ment

0.79

 forward

0.77

rers

0.75

ler

0.73

Activations Density 0.027%

instances where a significant action or decision is being reported on or discussed

No Comments

No Known Activations