INDEX

Explanations

references to subscribing or subscription-related actions

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

inas

-0.72

cule

-0.72

esan

-0.71

ball

-0.68

rans

-0.61

 Stephenson

-0.61

lda

-0.60

jas

-0.59

 Palm

-0.59

 impossibility

-0.59

POSITIVE LOGITS

Interstitial

0.86

 subscriptions

0.86

 subscribe

0.84

CHAT

0.81

scribe

0.81

 subscrib

0.75

 subscribers

0.74

ENC

0.74

 unsub

0.73

untarily

0.72

Activations Density 0.005%

references to subscribing or subscription-related actions

No Comments

No Known Activations