INDEX

Explanations

repeated phrases or words, particularly focusing on 'ga' and variations, indicating a search for patterns or emphasis in textual content

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

å¼¾

-0.16

TREE

-0.14

lez

-0.14

omain

-0.13

era

-0.13

celik

-0.13

inner

-0.13

GRAPH

-0.13

ãĥ³ãĥ

-0.13

cene

-0.13

POSITIVE LOGITS

iliz

0.18

 know

0.17

 known

0.16

tax

0.16

 Known

0.15

 taxpayers

0.15

ĽĦ

0.15

 Hazel

0.15

-under

0.15

 taxpayer

0.14

Activations Density 0.003%

repeated phrases or words, particularly focusing on 'ga' and variations, indicating a search for patterns or emphasis in textual content

No Comments

No Known Activations