INDEX

Explanations

technical terms related to research, development, and analysis

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

"},"

-0.79

Sund

-0.57

dr

-0.56

 Inher

-0.54

 merge

-0.53

 Rumble

-0.53

/"

-0.52

venge

-0.52

 fulfil

-0.51

 Himself

-0.51

POSITIVE LOGITS

,.

0.89

ebus

0.82

0.81

 importantly

0.72

adays

0.71

ãĢģ

0.71

,...

0.69

.,

0.67

acket

0.64

intendent

0.63

Activations Density 0.897%

technical terms related to research, development, and analysis

No Comments

No Known Activations