INDEX

Explanations

references to the word "hop" in various contexts

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

 bruising

-0.78

 Breed

-0.70

anguage

-0.70

UCT

-0.68

 Devils

-0.68

 Monstrous

-0.64

 Dynamics

-0.64

 behold

-0.63

IFE

-0.63

 Pats

-0.62

POSITIVE LOGITS

efully

1.72

eful

1.38

eless

1.07

hop

1.05

daq

1.04

yright

0.92

emaker

0.91

rog

0.90

roxy

0.89

ort

0.86

Activations Density 0.003%

references to the word "hop" in various contexts

No Comments

No Known Activations