INDEX

Explanations

references to collections or selections of items

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

uca

-0.76

inition

-0.75

atform

-0.74

Rating

-0.71

 teasp

-0.71

soType

-0.71

ilyn

-0.70

ittee

-0.68

umption

-0.67

otiation

-0.66

POSITIVE LOGITS

 sorts

1.09

 goodies

1.05

 essays

1.01

 photographs

0.96

 items

0.88

 artifacts

0.87

 images

0.87

 grievances

0.87

 anecdotes

0.85

 photos

0.82

Activations Density 0.065%

references to collections or selections of items

No Comments

No Known Activations