INDEX

Explanations

references to organizations or agencies with the initial "I" followed by a number

oai_token-act-pair · gpt-3.5-turbo

occurrences of the letter "I" as a standalone character

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 tongues

-0.67

 electromagnetic

-0.66

 optics

-0.65

 Cutter

-0.62

 Rebels

-0.61

terday

-0.60

 Bottom

-0.59

 membrane

-0.59

 ringing

-0.59

 Dill

-0.59

POSITIVE LOGITS

'm

1.29

've

1.05

verson

1.04

MAX

1.03

pec

1.02

AAF

1.01

ago

1.01

UC

1.00

CS

1.00

OA

1.00

Activations Density 0.205%

references to organizations or agencies with the initial "I" followed by a number

occurrences of the letter "I" as a standalone character

No Comments

No Known Activations