INDEX

Explanations

statements with emphasis or key points

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

 semblance

-0.76

 shroud

-0.74

creen

-0.73

 couch

-0.72

 guiActiveUnfocused

-0.69

 indo

-0.67

 detached

-0.67

 Marketable

-0.65

 playbook

-0.65

 board

-0.64

POSITIVE LOGITS

ª

1.39

ł

1.36

ĳ

1.23

¹

1.18

Ĳ

1.16

¡

1.12

Ĵ

1.11

ı

1.07

©

1.06

ľ

1.06

Activations Density 0.287%

statements with emphasis or key points

No Comments

No Known Activations