INDEX

Explanations

terms related to violent events, such as shootings and wars

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

gers

-0.70

ĪĴ

-0.69

thens

-0.69

ilk

-0.69

stal

-0.68

shire

-0.67

yssey

-0.67

BILITIES

-0.67

nos

-0.66

tis

-0.65

POSITIVE LOGITS

achusetts

1.48

achus

1.06

imil

0.86

acre

0.85

achu

0.82

 transit

0.79

aging

0.79

 exodus

0.78

eval

0.77

ages

0.77

Activations Density 0.671%

terms related to violent events, such as shootings and wars

No Comments

No Known Activations