INDEX

Explanations

code-related terms and constructs

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 14-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L14R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.14.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

yer

-0.15

ÏĩÎ½

-0.15

Î¼Î²

-0.15

eday

-0.15

Ð½ÑĸÑĪ

-0.14

.writerow

-0.14

Ø§ØŃ

-0.14

pir

-0.13

eller

-0.13

ubat

-0.13

POSITIVE LOGITS

 unin

0.16

onia

0.15

heck

0.15

ueur

0.15

uns

0.15

raft

0.14

bei

0.14

cm

0.14

 Brock

0.13

jan

0.13

Activations Density 0.000%

code-related terms and constructs

No Comments

No Known Activations