INDEX

Explanations

words related to legal cases or orders, specifically mentioning Seattle Times

oai_token-act-pair · gpt-3.5-turbo

the presence of specific sequences of characters or unusual formatting in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ASA

-0.77

ystem

-0.76

DAC

-0.74

ãĥ¼ãĥĨãĤ£

-0.74

roups

-0.72

 Faces

-0.69

Avg

-0.67

 Hear

-0.67

LAN

-0.67

-0.66

POSITIVE LOGITS

obi

0.74

lesi

0.73

 sealing

0.72

apor

0.70

reci

0.70

uting

0.69

OIL

0.68

Wik

0.67

à

0.67

onding

0.66

Activations Density 0.000%

words related to legal cases or orders, specifically mentioning Seattle Times

the presence of specific sequences of characters or unusual formatting in the text

No Comments

No Known Activations