INDEX

Explanations

names and references to prominent individuals and websites in the context of news and stories

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.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

elik

-0.16

-0.15

ayer

-0.15

pun

-0.15

orne

-0.15

DSL

-0.15

ynom

-0.15

uin

-0.15

-0.14

DDL

-0.14

POSITIVE LOGITS

enery

0.16

bau

0.15

ICA

0.15

ç£

0.15

EDI

0.15

ten

0.14

pes

0.14

uzzer

0.14

Äįer

0.14

erval

0.14

Activations Density 0.032%

names and references to prominent individuals and websites in the context of news and stories

No Comments

No Known Activations