INDEX

Explanations

references to encyclopedic or informative content

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.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

ietet

-0.16

 McCl

-0.16

istar

-0.15

shield

-0.15

rik

-0.15

wet

-0.14

holm

-0.14

ÃŃd

-0.14

reau

-0.14

ocode

-0.14

POSITIVE LOGITS

uki

0.17

illy

0.16

oze

0.16

wiki

0.15

expo

0.15

Pax

0.15

pr

0.15

ol

0.15

anol

0.15

olie

0.14

Activations Density 0.060%

references to encyclopedic or informative content

No Comments

No Known Activations