INDEX

Explanations

punctuation marks

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

 underestimated

-0.73

 preventive

-0.73

 charm

-0.73

 immersion

-0.72

 misunderstood

-0.70

 exha

-0.70

 interf

-0.69

 juggling

-0.69

 grounding

-0.68

 persecuted

-0.68

POSITIVE LOGITS

NET

1.11

taboola

1.08

htm

1.01

jpg

0.99

wav

0.97

gif

0.94

 âĶľâĶĢâĶĢ

0.93

html

0.93

\-

0.91

Net

0.89

Activations Density 0.425%

punctuation marks

No Comments

No Known Activations