INDEX

Explanations

numerical values associated with significant entities or measurements

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

rolet

-0.76

bery

-0.70

awaru

-0.68

icularly

-0.66

staking

-0.64

orically

-0.62

furt

-0.62

hold

-0.61

ilant

-0.61

oteric

-0.61

POSITIVE LOGITS

nd

0.91

Ħ¢

0.87

ÃĹ

0.86

XL

0.75

+,

0.74

Â¢

0.72

âĦ¢

0.71

rd

0.71

Aren

0.70

Activations Density 0.050%

numerical values associated with significant entities or measurements

No Comments

No Known Activations