INDEX

Explanations

sources or references indicated with the word "Read"

oai_token-act-pair · gpt-3.5-turbo

instances of the word "Read" indicating links or calls to action for more information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ascal

-0.74

ilogy

-0.68

TEXTURE

-0.66

aviour

-0.62

oult

-0.62

cffffcc

-0.61

avement

-0.60

 heel

-0.60

rouse

-0.60

 USSR

-0.59

POSITIVE LOGITS

Write

0.87

Read

0.86

ahead

0.83

 aloud

0.82

 Continued

0.82

just

0.80

gon

0.77

 Read

0.76

dress

0.74

çīĪ

0.73

Activations Density 0.018%

sources or references indicated with the word "Read"

instances of the word "Read" indicating links or calls to action for more information

No Comments

No Known Activations