INDEX

Explanations

superlative descriptors such as "largest," "biggest," and "most."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 7-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L7R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.7.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

ety

-0.18

elow

-0.16

oku

-0.15

roz

-0.15

TEL

-0.15

empor

-0.15

ament

-0.14

empre

-0.14

ideo

-0.14

odel

-0.14

POSITIVE LOGITS

pector

0.17

/world

0.16

 =============================================================================↵

0.15

ĸī

0.14

PRESS

0.14

ãĥĥãĥĪ

0.14

rist

0.14

ç«ĭãģ¦

0.14

DonaldTrump

0.14

chester

0.13

Activations Density 0.012%

superlative descriptors such as "largest," "biggest," and "most."

No Comments

No Known Activations