INDEX

Explanations

phrases related to the discussion or quoting of various perspectives or statements by different individuals

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

ãĥİ

-0.72

atible

-0.64

otin

-0.63

paralle

-0.62

Fit

-0.61

estine

-0.60

lor

-0.60

 ILCS

-0.59

Appearances

-0.59

imilar

-0.58

POSITIVE LOGITS

 sarcast

1.06

 bluntly

0.98

 rhet

0.94

 emphatically

0.81

 diplom

0.81

 aloud

0.76

0.75

 confidently

0.73

 bitterly

0.71

 omin

0.71

Activations Density 0.458%

phrases related to the discussion or quoting of various perspectives or statements by different individuals

No Comments

No Known Activations