INDEX

Explanations

phrases related to updates or future activities

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

 CLSID

-0.67

Maker

-0.59

 understatement

-0.59

VICE

-0.58

-0.57

 Abyss

-0.57

________

-0.57

________________________

-0.56

Â¯Â¯Â¯Â¯

-0.56

 Emanuel

-0.55

POSITIVE LOGITS

stairs

1.22

dates

1.17

grading

1.16

dating

1.15

grades

1.11

graded

1.08

rights

1.07

votes

1.01

raised

0.96

river

0.95

Activations Density 0.938%

phrases related to updates or future activities

No Comments

No Known Activations