INDEX

Explanations

terms related to significant damage or impactful events

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

ivan

-0.86

idas

-0.81

 Parables

-0.78

 Byrd

-0.78

Ã£

-0.77

Knight

-0.76

-|

-0.75

umer

-0.72

endants

-0.72

 Levi

-0.72

POSITIVE LOGITS

 health

1.02

 pathology

0.97

 stats

0.89

health

0.88

 physic

0.88

 quality

0.86

 healthcare

0.81

 psych

0.81

 clean

0.79

 stabil

0.79

Activations Density 0.084%

terms related to significant damage or impactful events

No Comments

No Known Activations