INDEX

Explanations

mismatch

np_max-act · gemini-2.0-flash

This neuron detects occurrences of the word “mismatch” (and its close variants) indicating a discrepancy.

oai_token-act-pair · o4-mini Triggered by @xinyanhu8

New Auto-Interp

Configuration

andyrdt/saes-llama-3.1-8b-instruct/resid_post_layer_11/trainer_1

Dataset (Dashboard)

Various

Features

131,072

Data Type

float32

Hook Name

blocks.11.hook_resid_post

Architecture

standard

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

(actor

-0.07

ji

-0.07

 rahat

-0.06

 cố

-0.06

 nghĩ

-0.06

 Bethesda

-0.06

calloc

-0.06

classname

-0.06

 activities

-0.06

 každé

-0.06

POSITIVE LOGITS

 mismatch

0.11

ismatch

0.08

 mism

0.08

Mismatch

0.08

 discrepancy

0.07

ins

0.07

inf

0.07

 alignment

0.07

盤

0.07

Dem

0.06

Activations Density 0.005%

mismatch

This neuron detects occurrences of the word “mismatch” (and its close variants) indicating a discrepancy.

No Comments

No Known Activations

mismatch

This neuron detects occurrences of the word “mismatch” (and its close variants) indicating a discrepancy.

No Comments

No Known Activations