INDEX

Explanations

names and acronyms related to specific entities or notable figures

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

ãĤ©

-0.92

zbek

-0.87

boa

-0.77

ngth

-0.77

iov

-0.75

ãĥ¼ãĥ«

-0.75

CVE

-0.73

%);

-0.71

hiba

-0.69

ãĥŃ

-0.69

POSITIVE LOGITS

aughs

0.94

entimes

0.93

OPLE

0.92

hedon

0.89

enburg

0.74

iers

0.74

uesday

0.74

ier

0.73

ypes

0.71

icult

0.70

Activations Density 0.024%

names and acronyms related to specific entities or notable figures

No Comments

No Known Activations