INDEX

Explanations

terms related to government and congressional proceedings

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

undi

-0.15

macen

-0.15

rega

-0.15

steen

-0.14

igli

-0.14

eri

-0.14

_license

-0.14

Rue

-0.14

 fiberglass

-0.14

Ð¼Ð¾Ð²

-0.14

POSITIVE LOGITS

 funding

0.23

 appropri

0.23

appropri

0.20

 appropriated

0.20

 Byrne

0.19

FY

0.19

FY

0.18

 rider

0.17

 riders

0.17

 Appropri

0.17

Activations Density 0.049%

terms related to government and congressional proceedings

No Comments

No Known Activations