INDEX

Explanations

words related to entities or organizations

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

å£«

-0.72

strap

-0.71

manship

-0.70

phrine

-0.70

 Archangel

-0.67

Ø©

-0.66

 hearts

-0.64

 Spears

-0.62

 tremend

-0.61

ned

-0.61

POSITIVE LOGITS

ropy

1.43

rance

1.28

itled

1.24

raction

1.18

rants

1.13

reprene

1.12

renched

1.06

race

1.03

ire

1.02

ourage

0.99

Activations Density 0.006%

words related to entities or organizations

No Comments

No Known Activations