INDEX

Explanations

occurrences of error handling in code

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ear

-0.15

elo

-0.15

qed

-0.14

IFT

-0.14

itive

-0.14

iris

-0.14

tob

-0.13

 Purs

-0.13

 ÄĳÃ¬nh

-0.13

Ä±k

-0.13

POSITIVE LOGITS

 Twin

0.17

:=

0.16

 Jaune

0.14

 Grove

0.14

 Blur

0.14

ormsg

0.14

>(()

0.13

ÑĢÐ°Ñģ

0.13

ugen

0.13

Yue

0.13

Activations Density 0.009%

occurrences of error handling in code

No Comments

No Known Activations