INDEX

Explanations

words related to expressing opinions or views

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

 consecut

-0.67

eri

-0.67

ammy

-0.66

trap

-0.66

vous

-0.65

amn

-0.65

esses

-0.64

ilant

-0.63

eff

-0.63

 Explos

-0.62

POSITIVE LOGITS

 views

1.03

 opinions

0.95

 viewpoints

0.88

finder

0.87

 opin

0.85

 opinion

0.84

topic

0.82

 viewpoint

0.82

»

0.82

 expressed

0.81

Activations Density 0.890%

words related to expressing opinions or views

No Comments

No Known Activations