INDEX

Explanations

references to feelings of guilt or acknowledgment of wrongdoing

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 9-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L9R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.9.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Jah

-0.17

OUCH

-0.14

oucher

-0.14

GRADE

-0.14

ç¬Ķ

-0.14

ru

-0.14

NH

-0.14

Lor

-0.14

asures

-0.13

 Verde

-0.13

POSITIVE LOGITS

ÑĤÐ¸Ð½

0.16

tar

0.16

ayÄ±p

0.15

owan

0.15

 horizontal

0.14

 bour

0.14

«a

0.14

CLU

0.14

cken

0.13

 Rewards

0.13

Activations Density 0.015%

references to feelings of guilt or acknowledgment of wrongdoing

No Comments

No Known Activations