INDEX

Explanations

phrases or sentences within brackets

oai_token-act-pair · gpt-3.5-turbo

instances of brackets or symbols indicating segmentation or grouping

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

 cones

-0.71

 intellig

-0.71

 equivalents

-0.71

 seams

-0.70

 poisoning

-0.69

 stagger

-0.69

 values

-0.68

ores

-0.67

 piping

-0.67

 transported

-0.67

POSITIVE LOGITS

...]

1.55

âĢ¦]

1.54

Pg

1.29

!]

1.22

](

1.07

AUT

1.06

actionDate

1.05

?]

1.04

paragraph

1.03

REDACTED

0.98

Activations Density 0.027%

phrases or sentences within brackets

instances of brackets or symbols indicating segmentation or grouping

No Comments

No Known Activations