INDEX

Explanations

words related to modification or transformation

oai_token-act-pair · gpt-3.5-turbo

the word "gold"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

Downloadha

-0.76

 sidx

-0.75

phrine

-0.75

 Marketable

-0.71

 Revolutionary

-0.68

OPLE

-0.68

 RESP

-0.66

 LINE

-0.65

 YEAR

-0.65

ãĥ´ãĤ¡

-0.63

POSITIVE LOGITS

ld

1.13

orf

1.07

roid

1.06

ynamic

0.92

irect

0.91

ouble

0.90

igger

0.84

ings

0.84

estone

0.82

sym

0.81

Activations Density 0.010%

words related to modification or transformation

the word "gold"

No Comments

No Known Activations