INDEX

Explanations

occurrences of the letter 'z' 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

ensis

-0.66

 Catalyst

-0.66

 peeled

-0.65

 Scots

-0.60

 pigeon

-0.59

ModLoader

-0.58

tailed

-0.58

 spoiled

-0.57

 Staples

-0.57

 mosqu

-0.57

POSITIVE LOGITS

ombie

0.94

ewski

0.94

ealous

0.85

immer

0.84

oker

0.82

ombies

0.81

cz

0.80

ulia

0.80

ulk

0.79

arnaev

0.78

Activations Density 0.010%

occurrences of the letter 'z' in various contexts

No Comments

No Known Activations