INDEX

Explanations

ratings or rankings

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

ĸļ

-0.83

skirts

-0.71

stem

-0.70

rote

-0.66

 Canaver

-0.65

wrong

-0.64

nown

-0.64

 Forth

-0.63

umb

-0.62

odcast

-0.62

POSITIVE LOGITS

ved

0.91

vation

0.88

iated

0.77

iation

0.74

 ratings

0.72

uated

0.72

iate

0.71

itect

0.70

osterone

0.70

IOR

0.69

Activations Density 0.004%

ratings or rankings

No Comments

No Known Activations