INDEX

Explanations

mentions of personal experiences and possessions

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

utan

-0.64

hov

-0.63

rities

-0.60

esson

-0.58

Lenin

-0.58

Created

-0.57

hire

-0.57

rior

-0.56

verages

-0.56

namese

-0.56

POSITIVE LOGITS

 doors

1.30

 Pandora

1.10

 door

1.04

 Doors

1.04

 valves

1.01

 gates

1.01

 portals

0.89

 backdoor

0.84

doors

0.80

 pores

0.79

Activations Density 0.048%

mentions of personal experiences and possessions

No Comments

No Known Activations