INDEX

Explanations

references to comparisons against "other" entities or subjects

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

isters

-0.80

staking

-0.74

acles

-0.72

stals

-0.69

ernels

-0.69

bda

-0.69

ensibly

-0.68

onics

-0.68

igers

-0.68

riages

-0.67

POSITIVE LOGITS

worldly

1.53

 aspect

1.08

 conceivable

0.97

 circumstance

0.96

 mammal

0.94

 thing

0.92

 person

0.90

 entity

0.89

 imaginable

0.88

 avenue

0.85

Activations Density 0.020%

references to comparisons against "other" entities or subjects

No Comments

No Known Activations