INDEX

Explanations

tweets with pictures

oai_token-act-pair · gpt-3.5-turbo

references to pictures

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

 footing

-0.68

NetMessage

-0.66

endi

-0.62

 silence

-0.61

 congress

-0.61

alle

-0.61

cffff

-0.61

 concentrate

-0.60

ifted

-0.58

 trave

-0.57

POSITIVE LOGITS

colo

1.10

amera

0.88

pic

0.86

 pics

0.85

twitter

0.80

://

0.77

img

0.76

ares

0.75

chrom

0.74

ographs

0.74

Activations Density 0.007%

tweets with pictures

references to pictures

No Comments

No Known Activations