INDEX

Explanations

references to procedures, evaluations, or assessments

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

Negative Logits

 Kelley

-0.17

NSSet

-0.15

é§

-0.14

-utils

-0.14

à¸ģà¸³

-0.14

igo

-0.14

 Direct

-0.14

plit

-0.13

owns

-0.13

 Personnel

-0.13

POSITIVE LOGITS

jar

0.16

soever

0.16

_fixture

0.15

pos

0.15

etched

0.14

Bars

0.14

Jar

0.14

jar

0.14

 analog

0.13

 Bars

0.13

Activations Density 0.014%

references to procedures, evaluations, or assessments

No Known Activations