INDEX

Explanations

words related to hidden or undisclosed information

oai_token-act-pair · gpt-3.5-turbo

occurrences of the prefix "un" indicating negation or reversal

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

anwhile

-1.01

å§«

-0.90

hyde

-0.87

ĺħ

-0.76

ŃĶ

-0.76

cium

-0.75

SHIP

-0.74

hetti

-0.72

phrine

-0.72

 briefs

-0.71

POSITIVE LOGITS

itled

1.14

ruly

1.14

ainted

1.04

rave

1.03

ested

1.00

ribut

0.98

ouch

0.98

enable

0.97

ired

0.96

ables

0.96

Activations Density 0.006%

words related to hidden or undisclosed information

occurrences of the prefix "un" indicating negation or reversal

No Comments

No Known Activations