INDEX

Explanations

numbers and measurements related terms

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "fig" and its variations, indicating a focus on numeric data presented in figure format

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

QUI

-0.75

 purity

-0.69

Ces

-0.66

 foul

-0.62

Gos

-0.62

 Rating

-0.62

 Wake

-0.61

 Scotch

-0.61

CHAT

-0.61

keyes

-0.60

POSITIVE LOGITS

uration

1.19

fig

1.06

lisher

0.99

lio

0.94

ãĥ¯

0.93

figure

0.92

thora

0.90

urations

0.90

roup

0.88

roups

0.87

Activations Density 0.017%

numbers and measurements related terms

occurrences of the word "fig" and its variations, indicating a focus on numeric data presented in figure format

No Comments

No Known Activations