INDEX

Explanations

references to states and state-related entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 13-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L13R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.13.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

teenth

-0.19

WARE

-0.17

angs

-0.17

onth

-0.16

bane

-0.16

rung

-0.15

ding

-0.15

thon

-0.15

edd

-0.14

_ONCE

-0.14

POSITIVE LOGITS

/local

0.21

-wide

0.20

wide

0.17

/reg

0.17

-level

0.16

-owned

0.16

-state

0.15

 Unidos

0.15

bound

0.15

opt

0.15

Activations Density 0.050%

references to states and state-related entities

No Comments

No Known Activations