INDEX

Explanations

references to the concept of "wrongness" or unacceptable behavior

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 21-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L21R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.21.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

-0.15

 Shepard

-0.15

ardy

-0.15

nul

-0.15

ery

-0.14

semb

-0.14

lias

-0.14

-stream

-0.14

pig

-0.14

POSITIVE LOGITS

erb

0.17

abel

0.15

_errno

0.15

-reset

0.15

ÑħÐ¸

0.15

AKER

0.14

ater

0.14

ATER

0.14

orf

0.14

ama

0.14

Activations Density 0.003%

references to the concept of "wrongness" or unacceptable behavior

No Comments

No Known Activations