INDEX

Explanations

phrases that indicate legislative actions or proposed bills

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

aukee

-0.67

 distur

-0.64

 frivol

-0.64

gor

-0.63

 fastball

-0.62

 bout

-0.62

-0.61

 bracelet

-0.61

ospace

-0.60

okia

-0.60

POSITIVE LOGITS

neath

1.37

lying

1.17

mining

1.09

graduate

1.07

statement

1.05

stand

1.03

cover

1.03

written

0.99

stood

0.97

reporting

0.95

Activations Density 0.021%

phrases that indicate legislative actions or proposed bills

No Comments

No Known Activations