INDEX

Explanations

phrases related to consequences or results happening as a result of specific events or actions

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

mate

-0.70

 cared

-0.69

tes

-0.69

wana

-0.69

 todd

-0.67

uded

-0.67

hatt

-0.67

inav

-0.66

alian

-0.64

itsch

-0.62

POSITIVE LOGITS

 confusion

0.95

 widespread

0.91

 riots

0.88

 outbreaks

0.88

 bloodshed

0.86

 deterioration

0.84

 refunds

0.84

 havoc

0.82

 predictable

0.81

 expulsion

0.80

Activations Density 1.405%

phrases related to consequences or results happening as a result of specific events or actions

No Comments

No Known Activations