INDEX

Explanations

references to specific events or actions taking place

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

heit

-0.72

FOR

-0.69

Maker

-0.66

âĺ

-0.65

ãĤ¤ãĥĪ

-0.64

 Located

-0.64

Pos

-0.62

ETS

-0.62

Area

-0.62

HO

-0.62

POSITIVE LOGITS

 breakup

1.16

 debacle

1.06

 onslaught

1.04

 hiatus

1.04

 announcement

1.03

 altercation

1.02

 collapse

1.00

 takeover

0.97

 arrival

0.97

 massacre

0.96

Activations Density 3.520%

references to specific events or actions taking place

No Comments

No Known Activations