INDEX

Explanations

mention of figures or references within a text

oai_token-act-pair · gpt-3.5-turbo

references to figures or illustrations 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 @ 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

ophobic

-0.65

 banned

-0.64

ocracy

-0.63

 dictionary

-0.62

 wars

-0.61

 admin

-0.60

 broadcast

-0.60

 supers

-0.59

 multicultural

-0.59

 classics

-0.58

POSITIVE LOGITS

Fig

3.97

Fig

3.08

FIG

2.11

fig

2.03

 Figure

2.01

fig

1.91

FIG

1.72

Figure

1.70

 Figures

1.68

 Table

1.35

Activations Density 0.028%

mention of figures or references within a text

references to figures or illustrations in the text

No Comments

No Known Activations