INDEX

Explanations

statements related to dishonesty or misinformation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 23-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L23R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.23.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

é¡Į

-0.16

 ragaz

-0.15

 propos

-0.14

jong

-0.14

 succesfully

-0.13

ä¸Ī

-0.13

 STRICT

-0.13

apat

-0.13

blas

-0.13

 Occasionally

-0.12

POSITIVE LOGITS

 nowhere

0.23

 documents

0.22

 document

0.18

documents

0.17

 emails

0.17

 public

0.17

 repeated

0.17

 independent

0.16

 repeatedly

0.16

 records

0.16

Activations Density 0.620%

statements related to dishonesty or misinformation

No Comments

No Known Activations

statements related to dishonesty or misinformation

No Comments

No Known Activations