INDEX

Explanations

objects located at the top of something

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "top" in various contexts

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

ellow

-0.75

ufact

-0.74

agra

-0.65

theless

-0.62

upon

-0.62

_-

-0.61

 Inqu

-0.61

 Leone

-0.60

gm

-0.60

 fellow

-0.59

POSITIVE LOGITS

most

1.18

liest

0.82

iary

0.78

 thereof

0.77

 tier

0.73

Bottom

0.69

bottom

0.69

 scorer

0.68

 rope

0.67

top

0.66

Activations Density 0.062%

objects located at the top of something

occurrences of the word "top" in various contexts

No Comments

No Known Activations