INDEX

Explanations

expressions related to political controversies

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

©¶æ¥µ

-0.84

Cot

-0.77

Wol

-0.77

 Dres

-0.75

KS

-0.74

 Phelps

-0.74

 Elves

-0.73

 Mend

-0.72

 Lauder

-0.72

 Frey

-0.70

POSITIVE LOGITS

tenance

1.22

withstanding

1.15

actly

1.13

usterity

1.07

actory

1.07

secut

1.06

ificantly

1.06

acted

1.04

pect

1.02

astic

1.01

Activations Density 1.037%

expressions related to political controversies

No Comments

No Known Activations