INDEX

Explanations

mentions of specific groups or categories within a broader context

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

nery

-0.70

ysc

-0.69

oldemort

-0.64

 prol

-0.62

ober

-0.62

irez

-0.61

idation

-0.60

 anytime

-0.59

idia

-0.58

adal

-0.58

POSITIVE LOGITS

st

0.95

Ĳ

0.88

Īè

0.87

Ĭ±

0.87

ĪĴ

0.84

stad

0.82

ī

0.79

Ĥª

0.79

among

0.76

eteen

0.75

Activations Density 1.317%

mentions of specific groups or categories within a broader context

No Comments

No Known Activations