INDEX

Explanations

mentions of professionals or people with specific expertise, potentially related to analysis or criticism

oai_token-act-pair · gpt-3.5-turbo

words related to analysis and analysts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 ferry

-0.68

 buckle

-0.64

 bite

-0.63

Pun

-0.63

Pen

-0.61

age

-0.59

 Hammer

-0.59

 brake

-0.59

 truck

-0.58

 Iron

-0.58

POSITIVE LOGITS

sts

4.94

stic

1.39

stals

1.35

zes

1.22

st

1.21

stal

1.14

zing

1.13

tic

1.07

ses

1.07

zed

1.04

Activations Density 0.011%

mentions of professionals or people with specific expertise, potentially related to analysis or criticism

words related to analysis and analysts

No Comments

No Known Activations