INDEX

Explanations

patterns, sequences, or organized structures

oai_token-act-pair · gpt-3.5-turbo

repeated references to patterns

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

ascular

-0.78

omez

-0.77

zona

-0.76

vez

-0.71

ille

-0.69

IVERS

-0.67

rican

-0.67

UGE

-0.67

UGH

-0.66

hiro

-0.66

POSITIVE LOGITS

 pattern

0.93

 patterns

0.91

 Pattern

0.91

pattern

0.90

gradient

0.87

 Patterns

0.87

ĸļ

0.86

Pattern

0.86

eering

0.84

atile

0.81

Activations Density 0.018%

patterns, sequences, or organized structures

repeated references to patterns

No Comments

No Known Activations