INDEX

Explanations

references to a governmental or administrative office

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

New Auto-Interp

Configuration

jbloom/GPT2-Small-OAI-v5-128k-attn-out-SAEs/v5_128k_layer_0

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

131,072

Data Type

torch.float32

Hook Name

blocks.0.hook_attn_out

Hook Layer

Architecture

standard

Context Size

Dataset

Skylion007/openwebtext

Activation Function

topk

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

achev

-1.00

wcs

-0.91

tymology

-0.90

culosis

-0.87

RH

-0.81

elong

-0.81

accompan

-0.80

olor

-0.80

ramer

-0.79

estern

-0.79

POSITIVE LOGITS

veyard

0.82

 Gaal

0.79

querade

0.78

 Reincarn

0.74

ctuary

0.70

 silence

0.69

quez

0.68

vine

0.68

Pokemon

0.68

nesses

0.66

Activations Density 0.025%

references to a governmental or administrative office

No Comments

No Known Activations