INDEX

Explanations

years or time-related keywords

oai_token-act-pair · gpt-3.5-turbo

references to specific time periods and events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

onica

-0.72

ounding

-0.72

ONG

-0.72

 srfAttach

-0.70

reetings

-0.69

ESE

-0.69

 Ahead

-0.68

ranging

-0.68

IOR

-0.67

ACY

-0.67

POSITIVE LOGITS

 withdrew

1.13

 chose

1.10

 refused

1.05

 forgot

1.04

 became

1.03

 opted

1.03

 undertook

1.03

 swore

1.03

 struggled

1.03

 threw

1.01

Activations Density 0.287%

years or time-related keywords

references to specific time periods and events

No Comments

No Known Activations