INDEX

Explanations

phrases related to reasoning or causality

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

wn

-0.74

lez

-0.74

ax

-0.70

robe

-0.70

agin

-0.69

mint

-0.67

nin

-0.66

Gas

-0.64

hal

-0.64

age

-0.64

POSITIVE LOGITS

 they

1.09

 there

0.91

 nobody

0.90

it

0.87

 THEY

0.85

 otherwise

0.83

we

0.81

 unlike

0.80

*/(

0.78

he

0.75

Activations Density 1.159%

phrases related to reasoning or causality

No Comments

No Known Activations