INDEX

Explanations

occurrences of error handling phrases 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 @ 6-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L6R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.6.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

avers

-0.14

andal

-0.14

exterity

-0.14

pector

-0.13

ccione

-0.13

æĲ

-0.13

ARNING

-0.13

 Hawkins

-0.13

aves

-0.13

avern

-0.13

POSITIVE LOGITS

raki

0.16

åı£

0.15

Dummy

0.15

 Rupert

0.15

utex

0.15

afc

0.14

ensibly

0.14

.bpm

0.13

dummy

0.13

 Dummy

0.13

Activations Density 0.004%

occurrences of error handling phrases in code

No Comments

No Known Activations