INDEX

Explanations

nouns referring to people, places, and things, often in their plural forms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 14-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L14R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.14.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

ieres

-0.17

ectors

-0.16

ietf

-0.16

isans

-0.15

ruptions

-0.15

ehicles

-0.15

MENTS

-0.15

minent

-0.14

ci

-0.14

ewe

-0.14

POSITIVE LOGITS

dum

0.17

rus

0.17

curity

0.16

acco

0.16

owing

0.16

 Huck

0.15

cond

0.15

two

0.15

few

0.15

quence

0.15

Activations Density 0.056%

nouns referring to people, places, and things, often in their plural forms

No Comments

No Known Activations