INDEX

Explanations

statements followed by factual corrections

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 24-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L24R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.24.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

acho

-0.15

 ageing

-0.14

athan

-0.14

ath

-0.14

utors

-0.14

ATH

-0.14

 Wire

-0.14

 marc

-0.14

 resc

-0.14

onte

-0.14

POSITIVE LOGITS

 claim

0.25

 Universal

0.24

 Welfare

0.24

 Claim

0.23

 welfare

0.23

claim

0.23

Universal

0.21

 claiming

0.21

 UNIVERS

0.20

Claim

0.20

Activations Density 0.030%

statements followed by factual corrections

No Comments

No Known Activations