INDEX

Explanations

negative sentiment or criticism regarding specific actions or decisions, often related to politics or economics

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

Administ

-0.64

ording

-0.63

Minor

-0.58

ĻĤ

-0.57

 Digest

-0.56

cknow

-0.54

 ende

-0.54

general

-0.53

 extent

-0.53

itty

-0.52

POSITIVE LOGITS

 onto

1.67

 into

1.46

into

1.23

 Into

1.16

 INTO

1.05

 overboard

0.97

 toward

0.94

izont

0.93

onto

0.92

 towards

0.86

Activations Density 2.137%

negative sentiment or criticism regarding specific actions or decisions, often related to politics or economics

No Comments

No Known Activations