INDEX

Explanations

phrases related to support, attention, and responses related to political or social situations

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

ovie

-0.65

igmat

-0.63

 Tale

-0.61

 sshd

-0.58

 Schmidt

-0.57

apixel

-0.55

Ri

-0.55

 lookout

-0.55

bey

-0.55

 constant

-0.54

POSITIVE LOGITS

fulness

0.97

flows

0.88

 emanating

0.83

killers

0.82

abilities

0.81

ternity

0.79

 from

0.78

worthy

0.77

 elsewhere

0.77

shed

0.76

Activations Density 1.828%

phrases related to support, attention, and responses related to political or social situations

No Comments

No Known Activations