INDEX

Explanations

phrases that indicate a change or guidance in movement or strategy

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

esters

-0.72

aqu

-0.68

itted

-0.66

oak

-0.66

asts

-0.65

 Byrne

-0.63

enos

-0.63

ammy

-0.62

Mini

-0.62

tein

-0.61

POSITIVE LOGITS

 direction

1.29

ality

1.15

 directions

1.07

finding

0.93

ward

0.92

ally

0.90

finder

0.87

 towards

0.86

 toward

0.85

ational

0.84

Activations Density 0.011%

phrases that indicate a change or guidance in movement or strategy

No Comments

No Known Activations