INDEX

Explanations

patterns resembling nested or structured code segments

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

rapped

-0.18

apur

-0.16

ĵåĲį

-0.16

gay

-0.16

ells

-0.15

pressions

-0.15

ookie

-0.15

inded

-0.15

å¸¯

-0.14

ple

-0.14

POSITIVE LOGITS

 Leone

0.15

#ad

0.15

upe

0.15

ĭ

0.14

atori

0.14

anuts

0.14

asal

0.14

dsp

0.13

//**↵

0.13

>",

0.13

Activations Density 0.006%

patterns resembling nested or structured code segments

No Comments

No Known Activations