INDEX

Explanations

references to previous positions or states

oai_token-act-pair · gpt-3.5-turbo

terms related to past and present states or positions

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

leon

-0.80

ARCH

-0.77

 Canaver

-0.77

Chip

-0.76

witz

-0.76

KER

-0.74

forge

-0.72

pour

-0.70

 Ridge

-0.69

ghai

-0.69

POSITIVE LOGITS

 counterparts

1.18

 arsenal

1.13

 selves

1.12

 holdings

1.05

 cousin

1.04

 counterpart

1.03

 surroundings

1.02

 repertoire

1.00

 predicament

0.98

 cousins

0.98

Activations Density 0.189%

references to previous positions or states

terms related to past and present states or positions

No Comments

No Known Activations