INDEX

Explanations

patterns and structures in programming and query-related contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 9-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L9R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.9.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

inski

-0.16

 Perr

-0.15

rase

-0.14

strup

-0.14

izzo

-0.14

rig

-0.13

uk

-0.13

assi

-0.13

andle

-0.13

POSITIVE LOGITS

’ÑıÐ·

0.15

gart

0.15

iams

0.15

erken

0.14

 =============================================================================↵

0.14

observ

0.14

izin

0.14

Observer

0.14

cala

0.14

disposing

0.13

Activations Density 0.174%

patterns and structures in programming and query-related contexts

No Comments

No Known Activations