INDEX

Explanations

programming-related structures and control flow elements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 11-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L11R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.11.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

raquo

-0.16

cete

-0.15

 ÙħØ´

-0.15

lernen

-0.15

umm

-0.14

íĨ¤

-0.14

ankan

-0.14

canf

-0.14

imdi

-0.14

ìłĦ

-0.14

POSITIVE LOGITS

nor

0.17

atura

0.15

anim

0.15

 Masc

0.15

omb

0.15

Agr

0.15

 pros

0.14

Vet

0.14

_mut

0.14

Eug

0.14

Activations Density 0.005%

programming-related structures and control flow elements

No Comments

No Known Activations