INDEX

Explanations

phrases related to opinions, statements, and discussions, especially in written form

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

ogether

-0.56

atics

-0.51

 beneficiaries

-0.51

azard

-0.50

 imported

-0.48

 broom

-0.46

 discriminating

-0.46

aceae

-0.45

 pend

-0.44

odge

-0.44

POSITIVE LOGITS

 remarks

0.77

 sarcast

0.73

 rhet

0.71

 keynote

0.67

Talking

0.66

 aloud

0.64

rompt

0.62

 yesterday

0.62

 interviewer

0.61

 announcing

0.61

Activations Density 0.647%

phrases related to opinions, statements, and discussions, especially in written form

No Comments

No Known Activations