INDEX

Explanations

phrases related to self-defense and justification in legal contexts

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

 Gloss

-0.81

olds

-0.68

 Floating

-0.66

terson

-0.65

 rides

-0.65

 sinks

-0.64

 Cake

-0.64

hesda

-0.63

flying

-0.63

 Revis

-0.62

POSITIVE LOGITS

 behalf

0.87

respons

0.85

 instinct

0.84

ifiable

0.80

 judgement

0.80

 ACTIONS

0.79

 respons

0.79

 behest

0.78

 impuls

0.78

acted

0.76

Activations Density 0.058%

phrases related to self-defense and justification in legal contexts

No Comments

No Known Activations