INDEX

Explanations

statements indicating difficulty or challenges

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 24-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L24R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.24.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

ht

-0.17

macen

-0.16

atics

-0.15

impan

-0.14

 Concern

-0.14

 stro

-0.13

abei

-0.13

gent

-0.13

 concerns

-0.13

stead

-0.13

POSITIVE LOGITS

 impossible

0.20

 difficult

0.18

/im

0.18

 harder

0.17

-hard

0.17

 Impossible

0.17

/exp

0.16

 sometimes

0.16

 Hard

0.16

Impossible

0.16

Activations Density 0.031%

statements indicating difficulty or challenges

No Comments

No Known Activations