INDEX

Explanations

references to structured lists or catalogues of information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

pector

-0.17

pawn

-0.15

vasion

-0.15

elage

-0.15

aba

-0.14

usra

-0.14

iage

-0.14

UBLE

-0.14

ifold

-0.14

ken

-0.14

POSITIVE LOGITS

0.15

è±Ĩ

0.15

 Dominion

0.14

ÄįnÃŃ

0.14

.cm

0.14

çħ§

0.14

ÏĦÏĮ

0.14

avir

0.13

 sÄ±ras

0.13

à¥įà¤Łà¤°

0.13

Activations Density 0.047%

references to structured lists or catalogues of information

No Comments

No Known Activations