INDEX

Explanations

phrases or terms related to decision-making and announcements

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

 instead

-0.74

 rall

-0.68

 rather

-0.68

rather

-0.66

 stayed

-0.65

instead

-0.62

pill

-0.62

 impro

-0.61

 seldom

-0.61

gul

-0.58

POSITIVE LOGITS

 specifics

0.94

 officially

0.90

nor

0.87

 definitively

0.85

 formally

0.82

 satisf

0.78

 definitive

0.78

yet

0.77

 conclusive

0.75

 finalized

0.74

Activations Density 0.113%

phrases or terms related to decision-making and announcements

No Comments

No Known Activations