INDEX

Explanations

references to various places or locations

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

DOS

-0.74

atorium

-0.71

CHAT

-0.70

sonian

-0.70

Progress

-0.70

FIN

-0.69

-0.68

oux

-0.67

ivation

-0.66

issance

-0.66

POSITIVE LOGITS

hare

0.97

 locations

0.95

 throughout

0.93

 frequ

0.93

hips

0.92

folk

0.86

ites

0.84

 around

0.81

pots

0.80

 inland

0.78

Activations Density 0.082%

references to various places or locations

No Comments

No Known Activations