INDEX

Explanations

phrases that inquire about actions or processes

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

wu

-0.16

asu

-0.14

ilogue

-0.14

abaj

-0.14

Ø¯Ø§ÙĨ

-0.14

ÑĽ

-0.14

 kodu

-0.13

é¥Ń

-0.13

ayers

-0.13

oÅ¾e

-0.13

POSITIVE LOGITS

yne

0.16

0.15

arda

0.14

RID

0.14

ards

0.14

DT

0.13

åł´

0.13

estre

0.13

Â¶

0.13

Activations Density 0.032%

phrases that inquire about actions or processes

No Comments

No Known Activations