INDEX

Explanations

phrases related to sufficiency or adequacy

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

kind

-0.71

CLASSIFIED

-0.70

ighth

-0.65

cember

-0.64

Ń·

-0.63

Ø±

-0.61

 renewal

-0.60

iaz

-0.60

icer

-0.59

Bah

-0.58

POSITIVE LOGITS

adena

0.70

yip

0.68

rehens

0.65

rehend

0.63

apesh

0.63

ritic

0.63

 imagination

0.62

 Already

0.62

ranged

0.61

 externalToEVAOnly

0.61

Activations Density 1.848%

phrases related to sufficiency or adequacy

No Comments

No Known Activations