INDEX

Explanations

actions related to adding or including something

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

NING

-0.68

yah

-0.66

Interested

-0.61

 Correspond

-0.59

grim

-0.58

ARE

-0.58

bis

-0.58

cies

-0.58

prototype

-0.58

Zel

-0.57

POSITIVE LOGITS

endum

1.20

 insult

1.07

itional

1.04

ressing

1.00

ictions

0.99

ition

0.97

itions

0.95

 thereto

0.93

resso

0.93

resses

0.93

Activations Density 1.706%

actions related to adding or including something

No Comments

No Known Activations