INDEX

Explanations

references to the idea of consistency or ongoing states

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

dor

-0.95

ocene

-0.90

lda

-0.85

ramid

-0.84

ahime

-0.81

endas

-0.78

md

-0.78

owitz

-0.77

baum

-0.75

eah

-0.75

POSITIVE LOGITS

motion

1.13

 vigilance

1.02

 reminders

0.92

 bombardment

0.92

 motion

0.92

 barrage

0.91

 reminder

0.90

 churn

0.88

 repetition

0.88

 bombard

0.85

Activations Density 0.015%

references to the idea of consistency or ongoing states

No Comments

No Known Activations