INDEX

Explanations

phrases that signal important instructions or warnings

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

 Ã¶yle

-0.16

ANA

-0.15

ana

-0.15

 áº¥y

-0.14

wan

-0.14

_SUITE

-0.13

 ê·¸ëŁ°

-0.13

 ê·¸ëłĩê²Į

-0.13

æ¥½

-0.13

 herself

-0.13

POSITIVE LOGITS

 there

0.22

 though

0.22

ascus

0.18

 always

0.17

 also

0.17

 Though

0.17

Though

0.17

erval

0.17

 although

0.17

0.16

Activations Density 0.048%

phrases that signal important instructions or warnings

No Comments

No Known Activations

phrases that signal important instructions or warnings

No Comments

No Known Activations